[moonman239 (OP)]

I know scientists measure the intensity of Earthquakes using either the moment magnitude scale or the Richter scale.  Let's say I have a device that contains an accelerometer, and I want to use the accelerometer to figure out how much "shake" an earthquake caused in an area.  How would I figure out, based on the data alone, how much that "shake" was?

I imagine such data would be useful to my county's emergency services office, because they could then figure out how much money and effort should be placed into preparing for a big earthquake.

[evan_au]

There have been proposals to use the accelerometers in many electronic phones (or air pressure in some newer mobile phones) to provide distributed sensing of seismic (or weather or explosion) events.

Most modern smartphones have a multi-axis accelerometer. Assuming the phone is sitting on a solid object, the acceleration of the phone will be proportional to the movement of the ground; the timing of earthquake arrival allows location of the epicenter. This can be translated into the ground acceleration of an earthquake as shown here.

It will take data collected from a region to be able to detect the location of the epicenter.

The static 1g acceleration of gravity must be subtracted from the sensor measurement, and any local vibrations such as passing trucks must be averaged, and considered as a threshold below which you don't consider ita potential earthquake. The Richter scale is logarithmic, with most earthquakes registering < 2 on this scale (very faint), and will be undetectable by this method.

The trick will be how to distinguish events like dropping the phone or jogging up the stairs from earthquake events. The sensor must remain powered-up to keep the sensor working, and this can shorten battery life; communication with nearby nodes also consumes power.

[moonman239 (OP)]

It will take data collected from a region to be able to detect the location of the epicenter.

The point of doing this is not to find the earthquake's center.  The point is to use the accelerometer data to figure out how much "shake" the earthquake caused in a given area.  Example:  Let's say said device uploads its data to the Internet.  One day, the sensor detects an earthquake whose center is a few miles away.  A friend of mine who lives elsewhere hears about the earthquake and wants to know if I may need emergency assistance or be dead, so after calling me and hearing my voicemail greeting he goes online to some Website that gets data from the data-collecting device.  The Website would help this friend determine if the earthquake could have caused me to be injured.

[evan_au]

The US Geological Survey already have a worldwide earthquake monitoring system, which reports the location of earthquakes to within a km or so, accessible here.

The advantage of crowdsourcing the earthquake intensity would be to provide a more fine-grained map of the impacted areas, to within a 100m or so.

It would probably cause concerns about "Big Brother", but if your smartphone and the mobile phone system recorded your last known location, that would probably help your friend find you and determine if you were in the impacted area (or, in the worst-case scenario, at least locate your body).

[Bored chemist]

I know scientists measure the intensity of Earthquakes using either the moment magnitude scale or the Richter scale.  Let's say I have a device that contains an accelerometer, and I want to use the accelerometer to figure out how much "shake" an earthquake caused in an area.  How would I figure out, based on the data alone, how much that "shake" was?

I imagine such data would be useful to my county's emergency services office, because they could then figure out how much money and effort should be placed into preparing for a big earthquake.

"How would I figure out, based on the data alone, how much that "shake" was?"
You probably can't, because it's difficult to tell a small nearby quake from a distant big one.
"I imagine such data would be useful to my county's emergency services office, because they could then figure out how much money and effort should be placed into preparing for a big earthquake."
Only if they have a time machine. By the time you have sensed a quake it's a bit late to start budgeting for preparations.

http://xkcd.com/723/

[evan_au]

The esiest way to help your country to prepare for earthquakes is to look at historical earthquake maps, like the one here; click to zoom in.

Most of the earthquakes cluster along tectonic plate boundaries.

There are some known limitations with this method:

• Historical records only go back a certain amount of time; if there were no written records of the last earthquake, it won't be marked on the map.
  
• Some sections of a plate boundary get "stuck", building up pressure until they release a massive earthquake (and possibly a large tsunami). This is what happened at Fukushima, and is (no doubt) happening at a number of other places around the world, like Seattle/Vancouver.
  
• To detect these "sleeping" earthquake zones, geologists need to do a lot more work, such as looking for silt layers far inland, or looking for debris from land slips on the continental shelf (such as occurs around Hawaii or the Canary Islands).
  
• Risk of an earthquake does not correlate directly with the risk of damage or injury. In areas that are thought to be "low risk", building codes are more relaxed, and so even a minor (unexpected) earthquake can cause major building collapses.
• Building on sand/gravel can produce problems during an earthquake, as the ground may turn liquid (such as happened in Mexico City, or more recently in the Christchurch earthquake). So the amount of damage can be much more severe than when building on rock foundations.